---
title: "Financial-Analysis"
output:
flexdashboard::flex_dashboard:
orientation: row
social: menu
source_code: embed
theme: lumen
vertical_layout: scroll
logo: india.png
favicon: t.png
---
```{r setup, include=FALSE,message=FALSE}
library(flexdashboard)
library(forecast)
library(quantmod)
library(dplyr)
library(PerformanceAnalytics)
library(rugarch)
library(plotly)
library(dygraphs)
library(prophet)
```
```{r,results='hide',cache=TRUE}
getSymbols("GS",from="2014-01-01",to="2020-09-10")
```
Data Table {data-icon="fa-table"}
=================================================
Time Series Plot {data-width=600}
------------------------------------------------
### Time Series Plot
```{r}
GS$GS.Close %>%
plot.xts(gsp,main="Goldman Sachs Inc.",
col="black",
bg = "pink")
```
### Data-Table
```{r}
GS %>%
data.frame() %>%
#tail(10)%>%
DT::datatable(colnames = c("Date","Open","High","Low","Close","Volume","Adjusted"))
```
Interactive {.tabset data-width=400}
--------------------------------------------------
### Line
```{r}
library(dygraphs)
final <- cbind(GS$GS.Open,GS$GS.High,GS$GS.Low,GS$GS.Close)
dygraph(final,
main = "Goldman Sach Price Repersentation") %>%
dySeries("GS.Open",label = "Open") %>%
dySeries("GS.High",label = "High") %>%
dySeries("GS.Low",label = "Low") %>%
dySeries("GS.Close",label = "Close") %>%
dyHighlight(highlightCircleSize = 3,highlightSeriesOpts = list(strokeWidth=3)) %>%
dyEvent("2020-01-15",label = "COVID",labelLoc = 'bottom') %>%
dyRangeSelector(fillColor = "violet",strokeColor = "pink",
dateWindow = c("2020-08-25","2020-09-09"))
```
### CandleStick
```{r}
G <- GS[,c("GS.Open","GS.High","GS.Low","GS.Close")]
colnames(G) <- c("Open","High","Low","Close")
#G$a <- NULL
#G$b <- NULL
#G$c <- NULL
dygraph(G) %>%
dyCandlestick() %>%
dyRangeSelector(fillColor = "lightblue",
dateWindow = c("2020-02-25","2020-09-09"))
```
Modeling {data-icon="fa-line-chart"}
===========================================================
Theoretical Explanation {.sidebar}
--------------------------------------------------------
### Time-Series ARMA Modeling
There are major assumption which \n assures good working of this model.
+ Stationary check:
Since P-value is
+ Constant Conditional Variance over the period of time.
+ Error is normally distributed.
DATA {data-height=100}
----------------------------------------------------------
### LAST PRICE
```{r}
library(flexdashboard)
valueBox(tail(GS$GS.Close,1),
caption = "Last Prevail Price",color = "green",
icon="fa-subscript")
##glue::glue("Price {lubridate::today()-1}")
```
### PREDICTED PRICE
```{r}
library(flexdashboard)
valueBox(round(predict(auto.arima(GS$GS.Close),1)$pred,4),
caption = "Auto-Arima Prediction",color = "lightblue",
icon="fa-superscript")
```
### Standard Error
```{r}
library(flexdashboard)
valueBox(round(predict(auto.arima(GS$GS.Close),1)$se,4),
caption = "Error Assumed",color = "grey",
icon="fa-user-secret")
```
GRAPH {.tabset .tabset-fade data-height=900}
----------------------------------------------------------
### ARIMA
```{r}
library(lubridate)
library(plotly)
library(ggplot2)
gsp <- GS$GS.Close
auto <- forecast::auto.arima(gsp)
df <- data.frame(yr=index(gsp),prc=gsp$GS.Close)
colnames(df) <- c("yr","prc")
at <- data.frame(yr=index(gsp),prc=auto$fitted)
colnames(at) <- c("yr","prc")
plotly::ggplotly(
ggplot(data=df,aes(x=yr,y=prc))+
geom_line(col="blue")+
geom_line(data = at,aes(col="red"))+
#geom_label(data = df,x=as.Date("2020-09-09"),y= 202.22,label="Actual")+
#geom_label(data=at,x=as.Date("2020-09-09"),y=203.2799,label="Prediction")+
xlab("")+ylab("")+ggthemes::theme_pander()+theme(legend.position = 'none')
)
```
### GARCH-GJR
```{r}
s <- ugarchspec(mean.model = list(armaOrder=c(1,1)),
variance.model = list(variance.model=list(model="gjrGARCH")),
distribution.model = "std"
)
m <- ugarchfit(spec = s, data = GS$GS.Close)
gdf <- data.frame(yr=index(GS),prc=m@fit$fitted.values)
colnames(gdf) <- c("yr","prc")
df <- data.frame(yr=index(gsp),prc=gsp$GS.Close)
colnames(df) <- c("yr","prc")
ggplotly(
ggplot(data=gdf, aes(x=yr, y=prc))+
geom_line(col="blue")+
geom_line(data = df, col="darkgreen")+
xlab("")+ylab("")+ggthemes::theme_pander()+
theme()
)
```
### PROPHET
```{r cache=TRUE}
library(prophet)
m <- data.frame(ds=index(gsp),y=gsp$GS.Close)
colnames(m) <- c("ds","y") #ds and y only work for prophet
mm <- prophet(m,weekly.seasonality = F)
future <- make_future_dataframe(mm,periods = 365)
forc <- predict(mm,future)
pp <- cbind(future,forc$yhat,forc$yhat_lower,forc$yhat_upper)
colnames(pp) <- c("yr","prc","lw","up")
pp <- as.xts(pp,order.by = pp$yr)
dygraph(pp) %>%
dySeries("yr",label = "Year") %>%
dySeries("prc",label = "Price") %>%
dySeries("lw",label = "low") %>%
dySeries("up",label = "Upper") %>%
#dySeries("close",label = "Actual") %>%
dyOptions(colors = c("lightblue","green","lightblue")) %>%
dyEvent("2020-09-09",label = "Prediction Start",labelLoc = "bottom") %>%
dyEvent("2020-01-15",label = "Covid-19",labelLoc = "bottom")
```
### TABLE
```{r}
s_final <- ugarchspec(mean.model = list(armaOrder=c(1,1)),
variance.model = list(model="gjrGARCH",garchOrder=c(1,1),
submodel=arima(GS$GS.Close,order = c(1,1,2))),
distribution.model = "sstd"
)
f_model <- ugarchfit(spec = s_final,data = GS$GS.Close)
setfixed(s_final) <- as.list(coef(f_model))
# f2014 <- ugarchforecast(data=GS$GS.Close["2014"],
# fitORspec = s_final,
# n.ahead = 252)
# f2021 <- ugarchforecast(data=GS$GS.Close["2020"],
# fitORspec = s_final,
# n.ahead = 252)
# par(mfrow=c(2,1))
# plot(sigma(f2014))
# plot(sigma(f2020))
path <- ugarchpath(spec = s_final,
m.sim = 3,
n.sim = 1*180,
rseed = 123)
#plot.zoo(fitted(path))
path@path$seriesSim %>%
as.data.frame() %>%
DT::datatable()
```
